Qwen3-Coder 30B-A3B Instruct
Smaller Qwen coder for efficient local agents and repo-level fixes
Shogo Agent Index
UnrankedNot ranked yet: 0 agentic benchmarks measured, and we need at least 2 plus an overall capability score. It will be ranked as results are published.
Agentic coding
—
estimated
Long-horizon autonomy
—
estimated
Knowledge-work agents
—
estimated
Computer use
—
estimated
Security & context learning
—
estimated
Price
- Input
- $0.18/M
- Output
- $0.70/M
- Cache read
- $0.014/M
- 3:1 blend
- $0.31/M
- Agent-effective
- $0.070/M
Cheapest host for agents: SiliconFlow at $0.080/M.
Agent-effective = 20 input tokens per output token, 85% cache hits. Why.
Agentic benchmarks
| Benchmark | Score | Normalised | Index weight | Harness / source |
|---|---|---|---|---|
| SWE-bench Verified | est. | 8% | — | |
| Terminal-Bench 2.0 | est. | 8% | — | |
| DeepSWE | est. | 6% | — | |
| FrontierSWE | est. | 5% | — | |
| CursorBench | est. | 4% | — | |
| GSO-Bench | est. | 4% | — | |
| METR Time Horizons | est. | 10% | — | |
| Vending-Bench 2 | est. | 6% | — | |
| PostTrainBench | est. | 4% | — | |
| APEX-Agents | est. | 8% | — | |
| GDPval | est. | 7% | — | |
| Remote Labor Index | est. | 5% | — | |
| DeepResearch Bench | est. | 5% | — | |
| OSWorld 2.0 | est. | 6% | — | |
| OSWorld | est. | 4% | — | |
| Cybench | est. | 5% | — | |
| CL-bench | est. | 5% | — |
API hosts
Live prices and uptime for every host serving Qwen3-Coder 30B-A3B Instruct via OpenRouter.
| Host | Input | Output | Cache read | Agent $/M | Uptime 24h | Quant |
|---|---|---|---|---|---|---|
| SiliconFlow | $0.070 | $0.28 | — | $0.080 | 99.3% | fp8 |
| Amazon Bedrock | $0.15 | $0.60 | — | $0.17 | 100.0% | — |